South Africa, the country with the most HIV-infected citizens globally, has the largest antiretroviral treatment (ART) program in the world. More than half of all South African adults have undergone voluntary counseling and testing (VCT) at least once. Unfortunately, these initiatives have not translated into early ART initiation. A substantial unmet need exists, with only 37% of the estimated 2 million HIV-infected ART-eligible individuals actually in care. The number of ART-eligible individuals will only increase, now that the government has adopted the new WHO guidelines, setting ART-eligibility at CD4 < 350 cells/mm3. To date, few studies have examined why millions of ART-eligible individuals in South Africa are not receiving treatment. We have identified a highly concerning phenomenon directly impacting ART uptake in South Africa. Over the course of 2009, 7,287 adults presented to a high-volume VCT site in Soweto. Ten percent (743) were found to be HIV-infected, and eligible to start immediate ART. We found that 148 (20%) of these clients refused to initiate ART upon learning their CD4 count, despite having a median CD4 of 110 cells/mm3. Though VCT is traditionally viewed as an entry point to treatment, these data underscore the importance of an unappreciated challenge in pre-ART care. To ensure the success of HIV treatment expansion, it will be essential to understand why individuals who qualify for ART choose not to initiate treatment. My goal with this K23 award is to identify modifiable determinants of ART-refusal, and design a targeted and informed intervention to counteract this phenomenon in Soweto, and prevent premature mortality in this vulnerable population. Using a mixed-methods approach, I will identify modifiable psycho-social determinants of ART-refusal by performing in-depth qualitative interviews with 40 ART-eligible adults presenting for VCT, and 10 healthcare providers, to inform category construction for a putative model based upon Anderson's Model of Healthcare Utilization. I will then examine correlates and outcomes of ART-refusal in 500 treatment-eligible adults, through survey implementation at the point of testing and for 6 months post-VCT. Finally, I will develop an intervention to promote HIV treatment acceptance among newly diagnosed HIV-infected individuals and pilot test it in a population of 100 treatment-eligible individuals. The proposed area of research addresses a critical area in pre-ART care, providing a theoretical and programmatic approach to improving ART uptake and acceptance among high-risk HIV-positive individuals in South Africa. The portfolio of research and training proposed for this K23 award will draw on my existing skills in HIV treatment and care, and global health program implementation. It will enable me to work with a multidisciplinary team of mentors across Harvard University and the University of Witwatersrand, and ultimately become a productive and influential independent investigator focused on improving ART acceptance.